A Dynamic Evaluation-Denoising Network for Motion Artifacts Removal from Single-Channel EEG.
Summary
This study introduces a novel Dynamic Evaluation Denoising Network (DED-Net) for enhanced brain-computer interface (BCI) signal processing. DED-Net effectively removes motion artifacts, improving signal quality and accuracy in neural remodeling and intent recognition tasks.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Brain-computer interfaces (BCIs) are crucial for rehabilitation and neural remodeling research.
- Existing BCI methods struggle with motion artifact generalization and denoising precision.
- These limitations hinder the practical application of BCIs.
Purpose of the Study:
- To develop an advanced denoising network for BCIs that addresses limitations in artifact handling.
- To improve the generalization ability and denoising precision of BCI signal processing.
Main Methods:
- Proposed a Dynamic Evaluation Denoising Network (DED-Net) integrating an evaluation model with cross-domain feature fusion.
- Employed dynamic selection of Bidirectional Long Short-Term Memory (Bi-LSTM) networks for artifact removal.
- Utilized EEGdenoiseNET for constructing a semi-simulated dataset for evaluation.
Main Results:
- DED-Net outperformed the state-of-the-art SDNet on a semi-simulated dataset, increasing SNR by 20.48% and CC by 3.15%.
- Achieved a signal-to-noise rate (SNR) of 6.0597 dB and a correlation coefficient (CC) of 95.28%.
- Demonstrated superior performance on real EEG data for intent recognition tasks, achieving 88.89% accuracy.
Conclusions:
- DED-Net offers superior performance in artifact detection, classification, and removal for BCI applications.
- The proposed method significantly enhances EEG signal reconstruction and intent recognition accuracy.
- DED-Net represents a significant advancement for practical BCI applications in rehabilitation research.
